Efficient Face Presentation Attack Detection | Muhammad Shahid Jabbar, PhD

Face presentation attack detection (face anti-spoofing) systems need to be both accurate under real-world conditions and cheap enough to run on constrained hardware. This project develops spatio-temporal deep learning frameworks with squeeze-and-excitation and Bahdanau attention mechanisms for robust detection, and multi-modal architectures that augment RGB input with synthetic depth modalities, achieving state-of-the-art performance with near-zero error rates on benchmark datasets.

A key thread is knowledge distillation: transferring what a heavier, depth-aware teacher model learns into a lightweight student model, so the resulting system stays deployable in resource-constrained and real-time settings (Jabbar et al., 2025). Related work has pushed this further toward IoT and edge deployment (Siddique et al., 2026). Complementary studies explore optical-flow and synthetic-depth augmentation (Jabbar & Khan, 2026), flow-based distillation (Jabbar et al., 2026), and attention-based spatio-temporal fusion for RGB-only input (Khan et al., 2026).

References

2026

IoT-J

RepGhostFAS: Lightweight and Efficient Face Anti-Spoofing for Internet-of-Things (IoT) and Edge Devices

Taha Hasan Masood Siddique, Shujaat Khan, Muhammad Shahid Jabbar, and 2 more authors

IEEE Internet of Things Journal, 2026

Bib PDF

@article{siddique2026repghostfas,
  title = {RepGhostFAS: Lightweight and Efficient Face Anti-Spoofing for Internet-of-Things (IoT) and Edge Devices},
  author = {Siddique, Taha Hasan Masood and Khan, Shujaat and Jabbar, Muhammad Shahid and Kumar, Akash and Huang, Kejie},
  journal = {IEEE Internet of Things Journal},
  year = {2026},
  publisher = {IEEE},
}

SPIE

Optical flow and synthetic depth augmentation for face presentation attack detection

Muhammad Shahid Jabbar and Shujaat Khan

In Real-time Processing of Image, Depth, and Video Information 2026, 2026

Bib PDF

@inproceedings{jabbar2026opticalflow,
  title = {Optical flow and synthetic depth augmentation for face presentation attack detection},
  author = {Jabbar, Muhammad Shahid and Khan, Shujaat},
  booktitle = {Real-time Processing of Image, Depth, and Video Information 2026},
  volume = {14087},
  pages = {1408705},
  year = {2026},
  organization = {SPIE},
}

Flow Augmentation and Knowledge Distillation for Lightweight Face Presentation Attack Detection

Muhammad Shahid Jabbar, Muhammad Sohail Ibrahim, Taha Hasan Masood Siddique, and 2 more authors

In 2026 IEEE 20th International Conference on Automatic Face and Gesture Recognition (FG), 2026

Bib PDF

@inproceedings{jabbar2026flowaugmentation,
  title = {Flow Augmentation and Knowledge Distillation for Lightweight Face Presentation Attack Detection},
  author = {Jabbar, Muhammad Shahid and Ibrahim, Muhammad Sohail and Siddique, Taha Hasan Masood and Huang, Kejie and Khan, Shujaat},
  booktitle = {2026 IEEE 20th International Conference on Automatic Face and Gesture Recognition (FG)},
  pages = {1--5},
  year = {2026},
  organization = {IEEE},
}

IWCMC

Attention-Based Spatio-Temporal Feature Fusion for Efficient RGB Face Presentation Attack Detection

Tazeen Khan, Muhammad Shahid Jabbar, Muhammad Sohail Ibrahim, and 2 more authors

In 2026 International Wireless Communications and Mobile Computing (IWCMC), 2026

Bib PDF

@inproceedings{khan2026attention,
  title = {Attention-Based Spatio-Temporal Feature Fusion for Efficient RGB Face Presentation Attack Detection},
  author = {Khan, Tazeen and Jabbar, Muhammad Shahid and Ibrahim, Muhammad Sohail and El-Maleh, Aiman Helmi and Khan, Shujaat},
  booktitle = {2026 International Wireless Communications and Mobile Computing (IWCMC)},
  pages = {1475--1480},
  year = {2026},
  organization = {IEEE},
}

2025

KBS

Knowledge Distillation with Predicted Depth for Robust and Lightweight Face Presentation Attack Detection

Muhammad Shahid Jabbar, Taha Hasan Masood Siddique, Kejie Huang, and 1 more author

Knowledge-Based Systems, 2025

Bib PDF

@article{jabbar2025knowledge,
  title = {Knowledge Distillation with Predicted Depth for Robust and Lightweight Face Presentation Attack Detection},
  author = {Jabbar, Muhammad Shahid and Siddique, Taha Hasan Masood and Huang, Kejie and Khan, Shujaat},
  journal = {Knowledge-Based Systems},
  pages = {114325},
  year = {2025},
  publisher = {Elsevier},
}